Dryspells and Minimum Air Temperatures Influence Rice Yields and their Forecast Uncertainties in Rainfed Systems

Abstract

Crop yield predictions on inter-seasonal to inter-annual horizons are subject to a diverse set of uncertainties associated with climate forecast scenarios. However, the uncertainties associated with climate-related parameters can be marginally controlled (reduced) over the growing season, i.e., from planting through harvest, leveraging a suite of climate forecast ensembles as forcings. In this study, we present a novel approach that combines a coupled hydrologic-crop modeling framework with probabilistic forecasts to characterize and reduce uncertainties in seasonal rice yields predictions. By weighting real-time (nowcast) and forecast climate forcings, the comprehensive framework accurately quantifies uncertainties associated with climate forecasts. At a provincial scale, the crop model extensively captured the uncertainties in yields whilst significantly complementing observations over the growing season. We observed that the spread of yield predictions gradually decreased over time (i.e., towards harvest) as subsequent timeframes incorporated a higher degree of present conditions/forcings and reduced reliance on forecasts. Furthermore, we investigated the information exchange between yields and hydrologic/drought variables over different time frames within the season. Notably, we found a higher synchronization of information transfer between yields, dryspells, and minimum air temperatures towards the end of season, indicating strong explicit links between these variables and crop yields. These outcomes have significant implications on crop yield forecasting and nowcasting, particularly in data-poor regions. By providing a better understanding of the uncertainties associated with seasonal climate forecasts and the interplay between hydrologic variables, drought conditions, and crop yields, this research can aid in improving decision-making processes related to agricultural planning, and risk management. Moreover, these insights can inform assessments of economic, social, and environmental impacts of drought in agricultural systems.